Repository navigation
Releases
v2.1.5
Full Changelog: v2.1.4...v2.1.5
v2.1.4
v2.1.3
v2.1.2
v2.1.1
v2.1.0
v2.0.0
Highlights of Tanuki v2.0.0
- Autonomous Identity Operator: Complete hands-free Active Directory diagnostics, Kerberos ticket lifecycle, and blast radius mapping.
- Quantitative Blast Radius Engine: Automated Tier 0/1/2 privilege mapping, RBCD delegation graph tracing, and Shadow Credentials inspection.
- Dual-Engine Protocol Parity: High-performance memory-safe Rust core (<0.96 ms latency) accompanied by a pure Python standard library runtime with zero external dependencies.
- In-Memory Clock Skew Self-Healing: Transparent KDC time offset compensation without requiring Linux CAP_SYS_TIME privileges.
- Deterministic Forensic Telemetry: Windows Security Event ID 4768 and 4769 correlation and NIST SP 800-88 single-pass memory purge.
v1.3.0
Overview
Tanuki v1.3.0 is a major protocol expansion delivering native Living-off-the-Domain (LotD) primitives and hardware-hardened Active Directory triage for Linux. Operating in pure unprivileged userland with zero external runtime dependencies, this release introduces native MS-PAC NDR Type 17 decoding, in-line Resource-Based Constrained Delegation (RBCD) DACL parsing, Shadow Credentials (msDS-KeyCredentialLink) analysis, RFC 6113 FAST armoring, and statically linked Musl Rust binaries.
Key Technical Enhancements
1. MS-PAC Bounded NDR Binary Decoder (tanuki pac)
- Native parsing of Microsoft Privilege Attribute Certificate (MS-PAC) structures directly from raw CCACHE ticket streams.
- Decoding of Buffer Type 17 (
PAC_ATTRIBUTES_INFO/ post-CVE-2024-26248 & CVE-2024-37085) evaluatingPAC_WAS_REQUESTEDandPAC_WAS_GIVEN_IMPLICITLYflags. - Complete NDR unmarshalling for
PAC_LOGON_INFO(Type 1),PAC_CLIENT_INFO(Type 10), and Kerberos ticket signatures without third-party dependencies.
2. In-Line RBCD Security Descriptor Parsing (tanuki ldap --query rbcd)
- Pure Python and Rust wire-level parser for
msDS-AllowedToActOnBehalfOfOtherIdentity. - Dissects relative NT Security Descriptors, extracting DACL control flags, Access Control Entries (ACEs), and trustee Security Identifiers (SIDs).
- Immediate identification of accounts holding delegation impersonation privileges against target domain objects.
3. Shadow Credentials & KeyCredentialLink Parser (tanuki shadow)
- Byte-level parsing of raw
msDS-KeyCredentialLinkbinary entries (both raw bytes and DN-Binary format). - Supports both Version 1 (ECC) and Version 2 (RSA) CNG public key structures (
BCRYPT_RSAPUBLIC_BLOB). - Extracts device ID GUIDs, raw public key material, key usage flags, and creation timestamps for shadow credential audit.
4. RFC 6113 FAST Armoring & Kerberos Hardening
- Support for Kerberos Flexible Authentication Secure Tunneling (FAST) via armored cache specification (
-T / --armor-cache). - Armored AS-REQ exchanges protect pre-authentication against offline dictionary attacks and ticket sniffing on unprivileged Linux workloads.
5. Static Musl Binaries & Dual-Engine Parity
- Compiled standalone static musl binaries for
x86_64-unknown-linux-muslandaarch64-unknown-linux-muslwith#![forbid(unsafe_code)]. - Zero shared library dependencies (
libc.so,libkrb5.sonot required for static binary operation). - Full functional parity between the Python standard library engine and the compiled Rust systems core.
6. Empirical Verification & Test Suite
- 411 automated unit and integration tests passing with 100% pass rate.
- Verified in live multi-OS virtual lab (Kali Linux 2024 against Windows Server 2022 Domain Controller).
- Verified dual light/dark mode assets and closed-loop Event ID 4768 / 5136 audit telemetry.
Distribution Packages
tanuki_ad-1.3.0-py3-none-any.whl: Universal pure-Python wheel (zero dependencies).tanuki_ad-1.3.0.tar.gz: Source distribution package.tanuki-x86_64-unknown-linux-musl: Standalone static Linux binary (x86_64).tanuki-aarch64-unknown-linux-musl: Standalone static Linux binary (ARM64).
Lineage & Pioneer Attributions
- Tim Brown (
@timb-machine): Linikatz post-exploitation and unprivileged Unix tradecraft pioneer. - Dietrich Gebert (
@dietrichayala): Ponytail decision ladder methodology pioneer.
v1.2.1
Overview
Tanuki v1.2.1 introduces deterministic semantic exit codes and machine-readable structured JSON error contracts across both Python and Rust engines. This update enables automated AI agent pipelines and DevOps operators to distinguish between policy guardrail stops, missing assets, parse corruptions, and usage errors without parsing conversational text.
Key Enhancements
1. Deterministic POSIX Semantic Exit Codes
Process exit codes are now standardized across both Python and Rust CLI entrypoints:
0(EXIT_SUCCESS): Command completed successfully or pre-flight diagnostic passed.1(EXIT_USAGE_ERROR): Invalid CLI options, missing required arguments, or unknown subcommands.2(EXIT_POLICY_STOP): Security policy or operational guardrail refusal (e.g. world-readable keytab permissions0644/0666, lowercase realm definitions, unconstrained delegation).3(EXIT_RESOURCE_MISSING): Target keytab file, SSSD socket, or credential cache not found on the filesystem.4(EXIT_PARSE_FAILURE): Corrupted ASN.1/RFC 4120 binary structures, invalid header magic, or truncated ticket streams.
2. Machine-Readable JSON Error Envelope
When invoked with --json, any error or policy refusal emits a structured JSON object to standard output:
{
"status": "REFUSED",
"reason_code": "INSECURE_PERMISSIONS",
"category": "POLICY_STOP",
"exit_code": 2,
"message": "Keytab file '/etc/krb5.keytab' has insecure permissions (0644). Secure mode requires 0600.",
"target": "/etc/krb5.keytab",
"details": "Mode 0644 allows non-privileged accounts to read Kerberos principal keys."
}For non-JSON invocation, errors continue to stream formatted diagnostic banners to standard error ([POLICY STOP], [RESOURCE MISSING], [PARSE FAILURE], [USAGE ERROR]).
3. Dual-Engine Parity
- Full functional parity maintained between the lightweight Python engine (
tanuki) and the compiled Rust systems core (tanuki-cli). - Both implementations share identical exit code integers, reason codes, and JSON schemas.
4. Empirical Verification & Packaging
- 315 automated unit and regression tests passing with 100% pass rate.
- Automated GitHub Actions CI workflow executed and verified across Python 3.10, 3.11, 3.12, container publishing to GitHub Packages (GHCR), and release package generation.
- Python wheels (
.whl) and source tarballs (.tar.gz) attached to this release.
Assets Attached
tanuki-1.2.1-py3-none-any.whl: Standard Python wheel.tanuki-1.2.1.tar.gz: Source distribution package.
Lineage & Pioneer Attributions
- Tim Brown (
@timb-machine): Linikatz post-exploitation tradecraft pioneer. - Dietrich Gebert (
@dietrichayala): Ponytail decision ladder methodology pioneer.
v1.2.0
Overview
Tanuki v1.2.0 expands the tool into a proactive, dual-use observability and Non-Human Identity (NHI) validation platform.
Key Additions
1. Pre-Flight Diagnostic Health Check (tanuki doctor)
- Runs in under 5 milliseconds with zero network packet emission.
- Verifies keytab permissions (
0600vs world-readable0644), realm uppercase consistency in/etc/krb5.conf, SSSD daemon socket status, and credential cache ticket lifetimes across file caches and Linux kernel keyrings (KEYRING:persistent:). - Supports human terminal output and structured
--jsonschema.
2. Dual-Use SOC Detection Telemetry
- Pairs tactical actions with defender detection rules: Auditd watch rules (
-w /etc/krb5.keytab -p r -k keytab_read), Windows Security Event IDs (4768, 4769, 4771), and Sigma/Falco signatures. - Terminal output displays
[BLUE TELEMETRY]alongside[TACTICAL CMD].
3. Non-Human Identity Token Exchange Validator (RFC 8693)
- Validates cloud workload JWT claims (Kubernetes ServiceAccount, AWS IAM Roles Anywhere, SPIFFE SVIDs) undergoing RFC 8693 OAuth 2.0 token exchange to Active Directory Kerberos tickets.
- Flags insecure wildcard audiences (
aud: "*") and non-ephemeral expiration lifetimes without external network dependencies.
4. Architecture & Documentation
- Native GitHub Mermaid architecture flowchart (
flowchart TD) and subsystem directory map. - Distroless/slim multi-stage Dockerfile and automated GitHub Actions workflow for container publishing to GitHub Container Registry (GHCR).
5. Empirical Verification
- 310 automated tests passing with 100% success rate across unit tests, pairwise matrices, and Tier 5 adversarial binary fuzzing.
Assets Attached
tanuki-1.2.0-py3-none-any.whl: Standard Python wheel.tanuki-1.2.0.tar.gz: Source distribution.
Installation
# Standard Python package
pip install tanuki-1.2.0-py3-none-any.whl
# Or install from source
pip install .
# 1-line installer (POSIX)
curl -sSL https://raw.githubusercontent.com/Mafifrizi/tanuki/main/install.sh | bash
# 1-line installer (Windows)
irm https://raw.githubusercontent.com/Mafifrizi/tanuki/main/install.ps1 | iexLineage & Pioneer Attributions
- Tim Brown (
@timb-machine): Linikatz pioneer. - Dietrich Gebert (
@dietrichayala): Ponytail methodology pioneer.